Heat pumps and the social cost of carbon

As shown in our general post about heat pump economics, the cost of heat pump conversions from natural gas heat, when computed per ton of carbon emissions reduction, will often be well above commonly accepted estimates for the social cost of carbon. Since gas is the dominant heating fuel, we may at some point need to revise the cost-benefit framework governing Mass Save.

Table of Contents

Defining the Social Cost of Carbon

The social cost of carbon is an economic concept used in benefit-cost analyses of proposals that may reduce carbon emissions. It is abbreviated as SCC or SC-GHG to include other green house gases.

The SC-GHG is the monetary value of the net harm to society associated with adding a small amount of that GHG [greenhouse gas] to the atmosphere in a given year. In principle, it includes the value of all climate change impacts, including (but not limited to) changes in net agricultural productivity, human health effects, property damage from increased flood risk natural disasters, disruption of energy systems, risk of conflict, environmental migration, and the value of ecosystem services. The SC-GHG, therefore, should reflect the societal value of reducing emissions of the gas in question by one ton.

Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990 Interagency Working Group on Social Cost of Greenhouse Gases, United States Government, page 9

The social cost of carbon is profoundly uncertain. One is trying to evaluate not only the future of global temperature and climate, but the impacts of changing climate on natural and human systems around the world. And then one has to put values on those impacts. Since those impacts happen well into the future, the SCC is highly dependent on what interest rate one uses to discount future harms. The mainstream solution has been to do repeated model runs to generate a statistical distribution of possible carbon cost/damage levels and then to evaluate the cost/damage distributions under several discount rates. (For background on the concept of “discount rates” see Wikipedia here.)

The graph below from the Technical Support Document of the Biden Administration’s Interagency Working Group on Social Cost of Greenhouse Gases shows distributions of alternative estimates of the 2020 social cost of carbon at each of three discount rates deemed reasonable: 2.5%, 3%, and 5%. These estimates are the estimates generated by the Obama administration with an update only for inflation.

Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990 Interagency Working Group on Social Cost of Greenhouse Gases, United States Government, page 28

For regulatory purposes, the social cost of carbon has to be a single number in any given year and many states and agencies have interpreted the federal analysis as recommending the use of the central estimate — $51, the average of the distribution under the 3% discount rate in 2020. However, the Biden working group’s TSD finds that (a) the Obama estimates do not cover all forms of damage adequately and (b) that the discount rates used may be too high (higher discount rates mean lower cost of emissions).

It is the IWG’s judgment that . . . the . . . SC-GHG estimates presented in this TSD likely underestimate societal damages from GHG emissions. . . . [Additionally, T]he IWG finds it appropriate as an interim recommendation that agencies may consider conducting additional sensitivity analysis using discount rates below 2.5%.

Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990, Interagency Working Group on Social Cost of Greenhouse Gases, United States Government, February 2021, pages 4, 21

They had to get an estimate up for use in regulatory proceedings and so adopted the work done under the Obama administration, but there is more work to do to reflect emerging science. The working group was charged with producing updated estimates by January 2022, but those have not emerged, perhaps in part as a result of litigation over the basic restoration of the Obama models. In a recently released report, the EPA is proposing a sharply increased estimate for the cost of carbon, but it is unclear how the EPA estimate relates to the still expected working group estimate.

SCC and heat pumps in Mass Save planning


Pursuant to Mass General Laws, Chapter 25, Section 19, the Department of Public Utilities has an obligation to make sure that Mass Save programs are cost-effective. Since our roadmap legislation of 2021, the department is obliged to consider the social value of carbon emissions reductions among the benefits in the benefit-cost analysis. In a cost-benefit analysis, a higher SCC increases the valuation of the benefits of any emission reduction. Increasing benefit valuations may make additional emissions reduction measures pass the cost-benefit screen.

In authorizing such programs, the department shall ensure that they are delivered in a cost-effective manner capturing all available efficiency opportunities, minimizing administrative costs to the fullest extent practicable; provided, however, that when determining cost-effectiveness, the calculation of program benefits shall include calculations of the social value of greenhouse gas emissions reductions, except in the cases of conversions from fossil fuel heating and cooling to fossil fuel heating and cooling . . . .

Mass General Laws, Chapter 25, Section 19 (emphasis added).

Note that the statute uses a positive term, “social value of greenhouse gas emissions reductions“; this term has been interpreted by all parties as cognate with the SCC.

In setting its social cost of carbon for use in evaluating measures to include its 2022-24 3 year plan, Mass Save initially (March 2021) received a recommendation to use an SCC of $128 per short ton. On October 12, 2021, the same consultant bumped their recommendation to $393 per short ton ($433 per metric ton). This value was then explicitly reflected in the term-sheet summary of negotiations about the plan concluded on October 25 and the proposed plan submitted to DPU on October 31.

Benefit Cost Ratio for Heat Pumps

Mass Save includes a number of types of benefits and costs in its benefit-cost analysis, but as the chart below shows, the SCC is an important component for heat pump conversions. The increase in social cost of carbon from $128 to $393 only moves a couple of types of heat pump conversions of gas furnaces to a benefit-cost ratio above one, but it raises all of the ratios substantially. Overall, it creates head room in the benefit cost analysis for the programs within which these measures are contained — “Residential Retail.” Without that head room, plan managers may have to manage consumer choices of incentives, etc., more tightly to assure a continuing positive benefit-cost level.

FuelHeat Pump Conversion MeasureBenefit Cost Ratio at $128 SCCBenefit Cost Ratio at $393 SCC
gasCentral Ducted Heat Pump Partially Displacing Existing Furnace, Gas0.390.78
gasCentral Ducted Heat Pump Fully Displacing Existing Furnace, Gas0.170.44
gasDMSHP* with Integrated Controls Partially Displacing Existing Boiler, Gas0.200.61
gasDMSHP* with Integrated Controls Fully Displacing Existing Boiler, Gas0.100.40
gasAir-to-Water Heat Pump displacing Existing Boiler, Gas0.551.12
gasClosed Loop GSHP** Replacing Furnace, Gas0.641.23
gasOpen Loop GSHP** Replacing Furnace, Gas1.252.34
oilCentral Heat Pump partially displacing Oil Heat1.592.37
oilCentral Heat Pump fully displacing Oil Heat1.792.57
oilMSHP* partially displacing Oil Heat1.622.43
oilMSHP* fully displacing Oil Heat1.622.37
oilAir-to-Water Heat Pump displacing Oil Heat2.113.17
oilClosed Loop GSHP** replacing Oil Heat2.173.24
oilOpen Loop GSHP** replacing Oil Heat3.725.59
Excerpted from attachment to National Grid response to Information Request DPU-Comm 8-2, Docket 21-124 and attachment. Get from original submission here.
*Ductless Minisplit heat pump. ** Ground Source HeatPump

The program administrators, in responses to DPU’s written questions, expressly acknowledged that some heat pump conversion “measures” — essentially the gas conversion measures listed above — are not cost-effective at the individual measure-level. National Grid explained its proposal to include some non-cost-effective measures within its “programs” in this response:

[T]hese measures deliver significant energy savings and greenhouse gas emissions reductions. These gas-to-electric offerings are part of a comprehensive effort by the PAs to meet the statewide GHG emissions reductions targets as set by the Secretary of the EEA on July 15, 2021. By creating an offering that allows all potential customers to participate, the PAs aim to create a program that best aligns with the Commonwealth’s policies and simplifies participation by all customers. There will be residents of the Commonwealth who seek to displace natural gas (partially or fully), even if it may not be in their short-term economic interests, for a variety of reasons, such as because they want to decrease their carbon footprint. There is evidence of adoption of said measures within the Commonwealth through the Massachusetts Clean Energy Center’s offerings and the Department of Energy Resource’s Home MVP pilot. The PAs design their programs to support all customers seeking to reduce energy and to meet all statutory obligations under the GCA and the Climate Act, and therefore the PAs believe these measures are appropriate and necessary to offer during the upcoming term. The PAs will continually review the measures’ effects on core initiative, program, and sector cost-effectiveness throughout the term.

National Grid response to Information Request DPU-Comm 5-3, Docket 21-124

This is a cogent response. From a marketing perspective, Mass Save’s approach makes sense — one wants to keep eligibility determination as simple as possible.

Litigation of the social cost of carbon

The unusual late increase in a fundamental program parameter, the social cost of carbon, raised concern at the DPU. DPU immediately signaled concern with a series of inquiries (see generally DPU’s early requests for information in Docket 21-124).

The briefs of the parties in support of the plan addressed the issue at length. The briefs of the parties leave no doubt that the goal of using a higher SCC was to render cost-effective as many measures as possible that might be needed to achieve the GHG reduction goals set by the Secretary of EoEEA for MassSave.

The Attorney General’s brief on this point is quite explicit. The brief refers to the “social discount rate.” The increase from $128 to $393 of the SCC was driven by a cut in the social discount rate from 2% to 1%. During the Statewide Plan hearings, DPU had raised the prospect of going back to 2% (so to $128 as an SCC).

During the Statewide Plan hearings, the Department pursued a line of questioning regarding the applied discount rate to reach the social value of GHG emissions reductions adopted in the 2022-2024 cost-effectiveness screening model. As part of this inquiry, the Department requested the Program Administrators re-run their plans as filed with a 2 percent social discount rate. An increase in the discount rate from 1 percent to 2 percent in the screening models and data tables would result in significantly lower benefits and lower cost-effectiveness compared to the filed Plan. Specifically, the 2022-2024 statewide electric benefits decline by nearly 3 billion dollars and the 2022-2024 statewide gas benefits decline by over 1 billion. Overall, increasing the social discount rate to 2 percent would result in a 29 percent decline in benefits. This decline in benefits would increase pressure on the Statewide Plan’s overall cost-effectiveness, [and] create the potential loss of measures for some Program Administrators that are critical to meeting GHG reductions, . . ..

Brief of the Attorney General, in Docket 21-124, p. 17 (emphasis added).

Similarly, the Acadia Center argued:

Using the appropriate [higher] social cost of carbon is an essential step for the PAs to develop a plan that is in compliance with the Climate Act – both its requirement to utilize the social cost of carbon in the BCR, and its requirement to meet the greenhouse gas reduction target set by the Secretary. The use of these updated [higher SCC] figures, as recommended by the update to the AESC study, are critical mechanisms to drive greenhouse gas emission reductions and “most appropriately reflected the Commonwealth’s policy goals on the urgency of avoiding the catastrophic intergenerational impacts of climate change.”

Brief of the Acadia Center in Docket 21-124, p. 17 (citations, notes, and stray punctuation omitted; emphasis added).

And to the same effect from Mass DOER:

[I]f the Social [Cost of Carbon] were revised downwards in the PA Plans as filed, the PAs have said that they would likely need to revisit the types and quantities of measures in the portfolio. Changing the types and quantities of measures risks a final plan that does not align with the priorities set forth in the Secretary’s GHG Goals letter or the Council’s priorities

Brief of MassDOER in Docket 21-124, p. 19 (emphasis added).

In its final order upon reviewing the three year plan at page 167 , the Department of Public Utilities, expressed disapproval of the process leading to the SCC increase and overruled the Mass Save Plan Administrators — ordering them to stay with the $128 SCC. As much as I support electrification, I found DPU’s concerns legitimate — legally, the “social value of greenhouse gas emissions reductions” has to be based on some kind of computation of value to society; it does not mean whatever it needs to mean to get the job done. See discussion below about an alternative statutory approach.

However, DPU then found that the proposed plan remained cost-effective and so approved it anyway:

The Department finds that, incorporating a $128 per short ton social value of GHG emissions reductions, the BCR [benefit-cost ratio] for all core initiatives, programs, and sectors for each Program Administrator’s Three-Year Plan remain cost effective and, therefore, no changes to the program designs included in the 2022-2024 Three-Year Plans are required (Tr. 2, at 286-289; RR-DPU-3, Att. A, Table IV.D.1).

Order upon reviewing the three year plan, page 176 (emphasis added).

The key words above are “initiatives, programs and sectors.” These are the terms used to outline the elements of the three-year plan. The table cited in the above quote from DPU is a recalculation of benefit-cost ratios using the lower $128 SCC which DPU requested from the plan administrators. It appears below for National Grid Gas. In the table below, the “sector” lines are grey; the “program” lines are purple; the “initiative” lines are white. At all of these rollup levels, the plan is cost-effective, although as noted, the initiatives include “measures” not shown which are not cost-effective. The DPU is reading the guiding statute above to require that the higher level groupings be cost-effective, but allowing that some programs’ component measures evaluated individually may not be. The statute is clear on this point — “For the purposes of reviewing cost effectiveness, programs shall be aggregated by sector. Any sector with a benefit cost ratio greater than 1.0 indicating benefits are greater than costs shall be considered cost-effective.” G.L. 25, s. 21(b)(3).

National Grid Response to DPU Records Request 3, Attachment A, Docket 21-124.

DPU specifically clarified that non-cost-effective gas to heat pump conversion measures are acceptable withing their larger program context in its Order on Motion for Clarification. But, by holding the plan administrators to the $128 SCC, DPU is leaving them less room for error.

How will it go next time?

For now, heat pump conversions from oil or resistance heat comprise most of the three-year plan’s electrification “strategic intervention.” Heat pump conversions from natural gas account for only 3.2% of the total planned 55,434 heat pump conversions. according to the Term Sheet. As a result, the numbers work — the plan “programs” are cost-effective.

Theoretically, the emphasis on oil could continue through 2030 — there are enough oil heated homes to accomplish our stated heat pump conversion goals. The Clean Energy and Climate Plan (relevant excerpts here) calls for an increase of 510,000 heat pumps by 2030 (including some new homes). According to the data here, there are a little under 700,000 oil heated homes now — enough to cover the targeted increase if very high market penetration could be achieved. However, there will be many oil homeowners that are not ready to convert. As part of the three year planning process, each of the utilities does “potential studies” (more on potential studies here) to differentiate between what is technically possible, what is economically reasonable, and what is practically achievable.

High technical and economic potential are attributed to HVAC energy optimization measures that completely or partially remove the fossil-fueled end-use load from a home. . . . [A]lthough [heat pump conversions] present a great technical opportunity for MMBtu savings, there are significant market barriers to customer adoption.

Massachusetts Energy Efficiency, Electrification, and Demand Response Potential Study for 2022-2024, Prepared for National Grid by Guidehouse, p. 23. Attachment to 2022-24 plan. See docket 21-124 in DPU file room.

So, it is likely that we will find a way to penetrate more deeply into the gas heated universe if we want to maintain a high annual volume of conversions. We will have to grapple with the fact that natural gas conversions do not work within the current benefit-cost framework. From the perspective of many consumers, the benefit-cost ratio is negative, regardless of incentives up to 100% of capital costs, since operating costs go up after conversion. And from the perspective of MassSave, for a typical heat-pump-from-gas conversion to exceed a threshold BCR of 1, the SCC would have to be almost $1,000. The graph below shows how high the social cost of carbon would need to be with current market pricing to achieve different benefit-cost ratios

Line extrapolated from the two BCRs at SCC of $128 and $393 in attachment to National Grid response to Information Request DPU-Comm 8-2, Docket 21-124 and attachment. The particular measure shown is “Central Ducted Heat Pump Fully Displacing Existing Furnace, Gas.” Get raw computations here.

An alternative to the social cost of carbon

Several distinguished economists have recently argued that the modeling approach used by the federal Interagency Working Group (a) is flawed and unreliable and (b) does not generate carbon “shadow prices” that are adequately high to support the investment needed to achieve stated climate goals.

Without delving into the arcane literature of “damage functions” and discount rates, suffice it to say that the uncertainty in looking so far to the future is irreducible. The federal regulatory framework may call for a quantified SCC and some economists may be able to form a conviction as to what the right SCC is, but their opinions will vary. Most people will never find a justification of any particular SCC to be compelling. Politicians will be able to vary it according to their ends. The federal SCC is likely to continue to lurch every few years as partisan control shifts.

The economists’ second point is also valid — the SCC just will not be high enough to support needed investment decisions within many regulatory frameworks, as shown above for our own MassSave framework.

We have committed to certain climate goals and it is possible to roughly assess the costs of achieving those goals. That is the exercise underlying our Clean Energy and Climate Plan. That plan has a finite number of moving parts. If we believe the plan is necessary, we need to price each of those parts, find funding for them, and move forward. We need to apply economic discipline to choose the most cost effective approaches, and some relative pricing model may help do that, but we will eventually need to dispense with benefit-cost screening using a carbon damage model. The question is not what theoretical future damages will be, but what we are willing to pay today to achieve our stated emission reduction goals. This reframing could address the MassSave benefit-cost ratio problem, but will consumers embrace conversions that raise their costs?

Will Brownsberger, November 2022

Return to heat pump document list

Leave a comment

Your email address will not be published. Required fields are marked *